Hans Keman. Handbook of Party Politics. Editor: Richard S Katz & William Crotty. 2006. Sage Publishing.
Introduction
Government, properly instituted, is a major impetus to economic growth, political development, and collective goods. Government, badly instituted, is a major font of poor economic performance, elitist privilege, and social waste. (Levi, 2002: pp. 54-5)
Governing society is one of the foremost topics for any student of representative democracy for it concerns directly the complexities of the politics-polity-policy triad (see Keman, 1997). This approach implies that party government is the irreducible core of any representative democracy because the political executive is constitutionally empowered to run the ‘affairs of the state’ based on a system of ‘checks and balances’ between the executive and legislature. This relationship differs across democratic polities, but at the end of the day it binds government in its capacities to act. In other words, the formal rules of the democratic game define the ‘room for maneuver’ of government and thus of the crucial actors making up government in all representative democracies: political parties.
The institutional context of government has been conducive to the development of informal rules, or conventions, to play the game in reality. These ‘rules of the game’ have emerged over time and define the parties’ actual scope for action. It is the interaction between parties that molds the pursuit of their main goals: policy-seeking and office-seeking. The former goal represents parties’ efforts to make government do what is in their interest and reflects their ideas about how society ought to be directed by means of public goods (e.g. socioeconomic policy-making, conducting foreign policy, etc.). The latter goal is to gain access to the decision-making arena (parliament and government) by competing with other parties. There is no policy-seeking behavior possible without being in office (seats in parliament or ministers in government). This type of behavior of parties and the resulting interaction between the executive and legislature is typical of parliamentary democracies (Strøm, 1990; Lijphart, 1999).
The core actors within systems of representative democracy, then, are political parties, i.e. the elected representatives having a ‘mandate’ to make policy choices and the ‘assignment’ to control government.Party government is the executive body responsible for policy-making and representing the top level within the polity. Hence, if one wishes to understand the working of the policy-making capacities of parliamentary party government, it is useful to introduce two concepts that indicate the ‘good governance’ of parliaments and governments (as expressed in the quote from Margaret Levi): whether or not government and parliament are sufficiently responsive to societal and political issues;2 and whether or not the policy choices made are indeed carried through (by parties) and effectively carried out (by government). This is what we call accountability. Both concepts allow for scrutinizing the procedural quality and material performance of representative government (Klingemann et al., 1994; Budge et al., 2002; Keman, 2002b).
In this contribution I shall elaborate on the act of governing by means of party government. Before scrutinizing this in more detail, in the next section I will introduce the debate on the position and role of party government that has dominated much of the literature on politics and government since the 1960s. From this debate a number of propositions have been derived which will be employed to dig deeper into the present ‘state of the art’ as regards party government. I shall first discuss the ‘history’ of this debate in qualitative and quantitative terms (see, for instance, Hibbs, 1992; Schmidt, 1996). This lays the foundation for the remainder of this chapter in which the various aspects of party government will be discussed.
Does Politics Matter—and do Parties in Particular?
As early as the 1960s, studies appeared explaining variations in public expenditure, either across nations or over time (Pryor, 1968; Wilensky 1975). On the one hand, it was claimed that ideological differences mattered in this respect, and on the other, it was stated that sociocultural and economic factors were decisive (for an overview, see Castles, 1981; Keman, 1993). This debate is still relevant for two reasons: First, if politics is not relevant for studying government, why bother about the role of parties in parliament and government? Second, if parties are not essential, why bother about the accountability and responsiveness of party government?
In retrospect this has been an important clash of views. For, if it can be demonstrated that electoral change between parties and a change in the party composition of government is related to changes in policy programs, then politics becomes an important factor not only in explaining policies, but also because this would imply that democratic governance makes a difference. An additional point of importance that made this debate relevant is that it implied a shift in focus within political science.
This debate made clear that the so-called ‘output’ (or public policy-making) ought to be taken into account. The debate on whether politics matters therefore had an important side effect: it made it clear that political science was not only about the relationships within the political system, but also about what party government produced for society (Lane and Ersson, 2000). The main attention of the participants in the debate has been on explaining the growth of ‘big government’ in the latter half of the 20th century (Wilensky, 1975; Castles, 1982; Hicks and Swank, 1992; Keman, 2002b). Those who claimed that ‘politics does not matter’ argued along three different lines, that state intervention is the result of: economic development and growth; structural social change; path-dependent and incremental trajectories. The counterarguments of political scientists were that, although economic resources and affluence are required to undertake public action, social needs and ideological preferences do shape the character and urgency of political demands. Yet, before one could seriously move on to explain how, to what extent and, in particular, why the patterns of public policy-making were also the result of politics and deliberate policy choices, these counterarguments had to be investigated. This has been conducive to a host of analyses that show that the ‘politics does not matter’ school cannot uphold its claim, nor is it empirically feasible to claim that parties and governments do not matter (for an overview, see Hibbs, 1992; Castles, 1998). Let us therefore turn now to the issue of how it is that parties matter.
Parties do Matter: But how?
Three different analytical clusters as regards the relationship between electoral representation of the citizen and the eventual pattern of governing represent the evolution of the debate: the impact of parties, in particular in relation to their policy-seeking behavior; the composition (type and color) of party government; and the form and organization of party government. Below we shall present the core elements of these three by means of the main theoretical propositions that represent the core of the ‘do politics and parties matter?’ literature.
The ‘Partisan Theory of Policy-Making’
Party democracy is characterized by the idea that the political process enhances the transformation of citizens’ preferences by means of party mediation (i.e. responsiveness) into policy choices and related governmental action (i.e. accountability). From this rather abstract point of departure one can infer a series of assumptions that allow for examining this process empirically.
First, the electorate has certain social and cultural characteristics that are conducive to having different preferences that will affect the direction and level of policy choices represented in party platforms (or manifestoes; for an empirical example, see Budge et al., 2001). This assumption has been elaborated by means of two propositions:
- The left versus right distribution of political parties across party systems influences policy choices (like the degree of state intervention in shaping society).
- The existence of organized class interests will influence the development of welfare state related policy choices (like the ‘butter versus guns’ discussion).
- Second, the role of parties is in part driven by their social constituency, and in part by their multi-functional organization: they are (and must be) policy-seeking and office-seeking (Strøm, 1990; Katz and Mair, 2002). The following propositions can be derived from this assumption:
- Differences between parties matter according to their size and representation in parliament as well as in government (this is the ‘office-seeking’ argument; see Laver and Schofield, 1990; Müller and Strøm, 1999).
- Parties in government pursue policies that are by and large compatible with their social constituency and policy program, i.e. the color of party government (this concerns the ‘policy-seeking’ argument; see Budge and Keman, 1990; Pennings, 1997).
Third, the capacity for policy-making of parties in government is dependent on the composition of government and the degree of executive dominance. This idea is reflected in two propositions:
- The type of government matters with respect to policy choices—it matters, for instance, whether there is a ‘single-party government,’ a ‘coalition government,’ or a ‘minority government’ (Gallagheret al., 2006; Woldendorp et al., 2000).
- The development of policy formation depends on how government works. Both the organizational make-up and the form of a party government will influence its viability and activities (van Roozendaal, 1997; Laver and Shepsle, 1996).
These assumptions and propositions of the ‘partisan theory’ are thus all characterized by the fact that actors, parties and governments, are considered to be crucial for explaining variations in governing and types of state intervention. This approach to the question of whether or not party politics matters has dominated the debate since the 1970s. Most analyses within the debate shared two characteristics. On the one hand, the dominant mode of analysis was of a comparative nature (and confined to the OECD world; see for instance, Castles, 1982). On the other hand, many studies tended to emphasize the policy performance of party government (e.g. Hibbs, 1992). Although this is a perfectly acceptable and viable research strategy, the danger is that ‘party government’ as such and the role of political parties in particular will disappear from the analysis. Keeping this caveat in mind, we shall now turn to empirical results of the debate on whether parties matter and then elaborate on the hypotheses central to this debate.
The Empirical Investigation of the Impact of Parties
The early empirical analyses almost always focused on the comparative influence of non-political factors (such as economic growth, age of the population, rates of unemployment and inflation: see Wilensky 1975; Cameron, 1978; Alt, 1985), on the one hand, and on political parties (in government or not) and the role of trade unions (Hibbs, 1977; Korpi, 1983; Esping-Andersen, 1985; Armingeon, 2002) on the other hand. In this type of comparison, political parties appeared less determinative than social and economic factors. However, the unexplained variation remained considerable (Keman, 2002b). Hence, there was still ample room for further analysis regarding parties’ influence on making policy choices. Subsequent research therefore focused on more ideological differences between parties and the impact of organized interests. These analyses revealed that the differences in terms of left versus right and a strong representation of, in particular, trade unions did indeed matter in making policy choices (Cameron, 1984; Laver and Budge, 1992; Huber et al., 1993; Cusack, 1997). Hence, parties and interest groups appear to matter and could account for the cross-national and cross-time variation in, for example, the development of the welfare state or the size of the public economy (see Castles, 1998).
More importantly, what came out of this empirically driven debate was the following:
- The presence of parties in government is more important than a party’s vote share or representation in parliament; hence incumbency seems to matter.
- The strength of the non-left parties in government is more relevant than the impact of—for example—social democrats as the main party of the left.
- The type and color of party government matter considerably as regards the actual policy outcomes for society.
In sum, the debate has led to the conclusion (ceteris paribus) that party government appears to be the pivot for studying partisan influence on policy-making. Yet one must consider in more detail the type, color and composition to assess why and how this is the case.
Party Government: Type, Color and Composition
The Type and Color of Party Government
After elections the distribution of parliamentary seats among the parties in competition is known and a government must be formed. The result of the formation leads to both the type and color of party government. These are by and large the result of a negotiation process (unless one party forms the government)—this will be examined in the next subsection. First we focus on types of government, using classifications derived from the composition of parliament.
The first distinction concerns the number of parties in government (one or more) and the second concerns a party having a majority or minority in parliament. In the past many political scientists argued that single-party majority government would be more efficient and effective in governing, for, so the argument goes, government is coherent (one party) and unified with respect to its policy program (see Duverger, 1968). This argument must be qualified, however, because it assumes that parties are unitary actors and governing only means (mechanically) executing policy priorities that are derived from party programs. As we know from empirical experience, this is not true: parties in government are not always unified in action, nor making policies consistent with their programs (an example of the former is the Liberal Democratic Party in Japan; examples of the latter are the Italian Democrazia Cristiana until the early 1990s and the French Gaullists).
A similar argument was made with respect to ‘minority’ government. First of all, such a government would hardly be capable of governing because it lacks sufficient parliamentary support. Second, due to lack of support this type of government would not last long. Yet, as for example Woldendorp et al. (2000: 86) found, 22% of all parliamentary governments in 48 democracies were ‘minority’ governments, lasting on average 440 days. Hence minority party government is not an exceptional phenomenon and its duration appears sufficient for effective governing. As Strøm (1984) has demonstrated (but see also Laver and Schofield, 1990; Warwick, 1994), minority governments do make policy. However, in contrast to other types of party government, minority governments obviously are dismissed more often due to conflict with the parliamentary majority. Hence, the distinction between the minority and majority type of government seems to be overdone and is less relevant than the distinction between ‘single-party’ and ‘multi-party’ government. The latter type is evidently always the product of coalition formation (see below).
A related distinction that turns out to be relevant for understanding the impact of parties is the color of government. This idea was introduced in the late 1980s (see Budge and Keman, 1990), and attempts to capture a more qualitative element. The variable reflects the presence of parties of the left, the center and the right according to their parliamentary strength and number of ministers in government. Hence, the color of party government indicates both the ideological tendency (i.e. centrality) of a government and the numerical weight (dominance) of the participating parties (see Laver and Shepsle, 1996; van Roozendaal, 1997). The idea behind this classification of governments has been to demonstrate not merely that parties do matter, but that the composition of government matters even more (van Roozendaal, 1992; De Winter, 2002). And, as we observed earlier, party differences do matter as regards policy-seeking behavior. We argue therefore that for the understanding of governing by parties in representative democracies, the type and color of government are indispensable features to take into account.
The Formation of Party Government
Apart from single-party government, the formation of governments is a complex process which in some cases (e.g. Belgium and the Netherlands) can take 3-6 months (Keman, 2002a). This is understandable. As we discussed above, it is not merely a matter of gaining governmental power (i.e. office-seeking), but of doing so in such a way that the collective choices as regards policy-making are as close as possible to the parties’ interests (i.e. policy-seeking).
Much theory on government formation has been developed over time, and a large number of overviews have been published (see Lijphart, 1984; Budge and Keman, 1990; Laver and Schofield, 1990; Müller and Strøm, 1999). The basic idea of most of these theories is that mutual collaboration between parties in government depends on necessity (to form a majority) and familiarity (to be able to cooperate). In addition, one can distinguish two ‘schools’ of thought that explain the eventual outcome of the negotiation process between parties. On the one hand, there is the office-seeking school that assumes that self-interested behavior drives all actors. On the other hand, there is the policy-seeking school where coalescence and cooperation emanate from (more and less) shared values and policy choices.
The office-seeking approach argues that numerical conditions predominate and determine the outcome, in particular that the number of parties in government is never more than is necessary for a parliamentary majority. This is called ‘minimal winning coalition government.’ Obviously this approach disregards ideological party differences as well as other types of social and political animosities that do exist in reality (e.g. cleavages in consociational democracies or class ideology in Scandinavia). Empirically, however, many coalitions formed are not minimal winning (Woldendorp et al., 2000).
The policy-seeking ‘school’ attempts to account for this lack of ‘reality’ and introduces the feature of party distances (in terms of left versus right or shared policy aims), which are based on divisions within party systems. The bigger the ideological gap between parties, the less likely these parties will form a government together. This assumption has induced new types of government: minimal winning (but) connected coalitions, minimal range coalitions, and policy viable coalitions. Yet this is basically an extension of the minimal winning coalition approach: the assumption remains that the smaller the number of parties, the happier the coalition will be. The genuine policy-seeking approach, however, leaves behind the numerical dimension and even, to some extent, the majority principle as a prerequisite for forming a viable party government. The formation process can then lead to types of party government that are ‘broad’ (cf. Lijphart, 1984) or ‘oversized’ (cf. Budge and Keman, 1990): there are more parties in government than strictly necessary (thus violating the minimum winning principle). In addition, this ‘school’ claims to explain the formation of durable minority governments. The argument is that the other parties in parliament have in common that they dislike each other more than the party (or parties) in government. As a result they are prepared to accept a minority government until they find a suitable alternative combination (Strøm, 1984). A final consideration within this approach is that particular circumstances (a ‘crisis’) or institutions (e.g. needing a qualified majority) may well be conducive to the formation and maintenance of a government that cannot be explained by policy- or office-seeking motives alone (e.g. in Belgium in the process of developing towards a federal polity, or in Switzerland to consolidate unity at the federal level).
More recently, government formation theory has tended to focus more on the process of formation and its ramifications for government composition. In addition to the color of party government, more attention is paid to the conditions under which parties negotiate (Laver and Budge, 1992; Müller and Strøm, 1999). An important feature of this process is ‘government agreement,’ on the one hand, and negotiating ‘portfolio distribution’ among coalition parties, on the other (Budge and Keman, 1990; Laver and Shepsle, 1994; De Winter, 2002). Government agreements concern the policy program the parties will pursue in government and the distribution of ministerial portfolios. In some cases these agreements are quite specific (and lengthy, for instance in the Benelux countries: Keman, 2002a). The agreement serves the purpose of binding all parties in government so as to preserve its policy viability. Often this implies a rather monistic relationship with parliament, which tends to become a powerless institution. Portfolio distribution is often characterized by the rule of proportionality: the relative size of the governing parties in parliament correlates highly with the number of ministries acquired (see Budge and Keman, 1990; Woldendorp et al., 2000). Hence, the portfolio distribution reflects the color of government by distributing the ministries mainly according to the policy preferences of participating parties (Laver and Budge, 1992; Müller and Strøm, 1999).
There are many theoretical explanations with respect to the formation of party government. Yet at the end of the day it is a process that is mainly directed by the organization and working of the party system with respect to party competition and programmatic differences. In addition, special features such as government agreements and conventions play their part. These features differ considerably across representative democracies.
The average duration of party government is below 2 years (Luxembourg, Spain, and the UK are on top, Greece, France and Italy are quite below the average). This means that party governments of whatever type are less durable than many think5. The types of government resulting from the formation process do vary, but most polities score between 2 and 3.5. This means that in many representative democracies the tendency is to form a minimal winning coalition or a surplus or oversized party government. The exceptions are the polities belonging to the Westminster type of democracy (see: Lijphart, 1984): Australia, Canada, New Zealand, and the United Kingdom. On the other hand there are those polities where a tendency towards minority government can be observed: Denmark, Finland, Italy and Poland. This is also reflected in their average support of government in parliament, which is below 50% (also in Norway and Sweden where minority and minimal winning coalition governments alternate). Finally the color of government, the proxy for the policy-seeking tendency of a party government, varies considerably: in a number of countries the average value is below 2.0 (i.e. rightwing or center right), but only in Sweden can the opposite be noticed (4.1). Yet, most noteworthy is the fact that more than one-quarter of all governments can be considered as centrist. Hence, type and color of party government do vary across and within nations and seem to matter in terms of government formation. Parties are apparently concerned about which parties they share power with (in a coalition) and how this may work out with respect to governing. This is also dependent on the ‘form’ of cabinet government.
Forms of Cabinet Government: Hierarchy and Collegiality
Governing in a parliamentary system depends very much on the balance between prime minister and ministers. How the executive is organized often implies a tension between collegiality and hierarchy, between a preeminent chief minister and a ministerial college of political equals.
Presently, it appears that the office of prime minister is acquiring more weight in the cabinet, even in quite egalitarian cabinet systems. It is argued that the increasing role of the media and thus of the prime minister as spokesman for government as a whole as well as the increasing importance of international meetings of government leaders have caused this. This development can be observed in particular in Western Europe due to the process of political integration of the European Union. Increasingly the final decision-making of the EU takes place via transnational bodies involving the national heads of government. At the same time there is an increased need for concertation and coordination of domestic policy formation. Nevertheless, the prime minister remains a ‘first among equals’ in a collegial cabinet government, since the principle of collegial decision-making is (still) predominant.
This recent development also implies a change with respect to the role of chairmen or leaders of parties in some countries. Especially in fragile coalition situations, these persons were often not included in government but remained in parliament, in part because they functioned as a ‘chief whip.’ Yet, in contrast to their Anglo-Saxon counterparts, these continental party leaders to a large extent control both the parliamentary party and their ministers in government (this has been the case in many consociational democracies: see Daalder, 1987). More recently, ‘other’ parties have tried to exclude dominant politicians from the coalition (for instance, in Austria and Italy). Finally, it ought to be pointed out that ministers, especially in ministerially organized cabinets, do not always honor their party mandate. Of course, this type of behavior can jeopardize the stability of government (Blondel and Thiébault, 1991).
The principle of ‘collegiality’ involves not only equality in rank-and-file within government, but also the idea that all decisions are made collectively. A minister who has been outvoted has no right to go public and to distance him or herself, but rather must share collective responsibility with the whole cabinet vis-ă-vis parliament. If not, then the minister is expected to resign. This convention is becoming rare, however, since in most systems nowadays, dissent more often than not means that the cabinet government as a whole resigns (De Winter, 2002). With a coalition government this is virtually a fixed, if informal, rule. The reason is that the parties in government will not allow the upset of the delicate inter-party balance established, and reflected in the portfolio distribution among the participating parties. This type of organization of government is almost exclusively Western European and is typical of the slow process of democratization in the 19th and 20th centuries, especially in constitutional monarchies. Two other types of cabinet government have evolved over time, however: prime ministerial cabinets, on the one hand, and ministerial governance, on the other.
Prime ministerial cabinets have developed in most Anglo-Saxon countries, where, due to the ‘first-past-the-post’ electoral system, there is (almost) always a majority party in parliament. Hence, this party forms the government and the leader is in a position to appoint and dismiss ministers. The United Kingdom and Canada are typical examples of this type of single-party government. Prime ministerial government also exists, however, in some parliamentary systems where a coalition is necessary to govern. Here the prime minister derives his or her dominant position from the formal relations between the executive and legislature: the prime minister is often less vulnerable because of the ‘constructive vote of no confidence,’ meaning that such a motion is only allowed if and when there is an alternative prime minister with a parliamentary majority (this principle exists, for example, in Germany and Spain). In this type of cabinet government it is the ‘chancellor’ who deals with parliament and with the individual ministers. In a sense, the chancellor is the ‘conductor’ and supervisor with respect to policy coordination. Although the role of the prime minister appears to be quite dominant, it must be noted that the stability of this type of cabinet government depends on the unity and homogeneity of the governing party. If there are strong rival factions, internal conflict may well lead to the replacement of the prime minister or of dominant ministers (this happens more often than not in Japan, but also occurs in the United Kingdom—Thatcher in 1990—and Germany -Brandt in 1974). Hence prime ministerial party governments are more hierarchically organized than collegial coalition government but not always more homogeneous.
Finally, there is the ministerial cabinet government. In this case the institutions are not in place to induce collegial behavior between the ministers, nor has the prime minister sufficient powers to act as a ‘supremo.’ Each and every minister is responsible for his or her policy area and, consequently, there is less policy coordination. In fact, the prime minister is basically a power broker who is involved in two arenas: within government and vis-ă-vis parliament. This form of party government can be found in Belgium and Italy (De Winter, 2002; Laver and Shepsle, 1996). It should not come as a surprise that ministerial cabinet governments are less enduring than others and are considered to be less efficient in decision-making compared with other government types.
The division of responsibilities is also a major feature of the organization of what I will call dual cabinet government (Weaver and Rockman, 1993). This type of government, inspired by the phenomenon of ‘semi-presidentialism,’ is based on a division of responsibilities, but here between the head of government and the head of state. In organizational terms it means that the decision-making powers are shared, while the implementation of policies as well as accountability vis-ă-vis parliament rest solely with the cabinet. Hence, the relationship within government is neither hierarchical nor exactly collegial in nature. In particular, the prime minister is in a delicate position: dependent on the president for a number of matters (often foreign affairs and defense) and responsible to parliament. It goes almost without saying that this form of government can be rather problematic, or at least is restricted in its actions. Dual cabinet government is further complicated by the fact that many of these are coalition governments or, to make things worse, the president is confronted with a hostile parliamentary majority (i.e. ‘cohabitation’ and ‘divided government’). This can be conducive to deadlock in decision-making and gridlock in policy implementation. Typical cases are France, and until recently Finland and Portugal. Yet, despite these drawbacks it can be noted that this type of organization of governing has been a model for many of the recently democratized countries in Central and Eastern Europe (Elgie, 1999).
Obviously the collegial form of government is the most prevalent type. This is quite understandable given the high number of parliamentary regimes in our universe of analysis. Nevertheless the more hierarchically organized forms together outnumber the collegial type. What also should be noted is that four out of the seven polities with prime ministerial organization are those of the established democracies that have experienced an autocratic regime in the 20th century. Another institutional legacy is the ‘constructive vote of confidence’ that is required in Germany and Spain to dismiss government. The other hierarchical cases are Anglo-Saxon, having inherited the Westminster type of political system (apart from Canada, where the central government has a distinctive role within its federal constitution: Braun et al., 2002).
The dual form of cabinet government goes, of course, together with variations of semi-presidentialism. Israel is the outlier in this respect, but belonged to this category due to the rule that the prime minister was directly elected. Hence both the head of government and the head of state are more or less independent of parliamentary intervention. This is also the case in South Africa, albeit that one person is head of both state and government. Dualist party governments thus have in common that heads of state and of government share political responsibilities. This makes this form different from the pure prime ministerial form.
The most prevalent form of party government in representative democracies is thus collegial cabinet government. It occurs in almost one-third of all countries under review here. What they have in common is that these countries are characterized by a multi-party system and thus by coalition government. The collegial principle can be considered as an institutional guarantee for the participating parties: on the one hand, it implies veto power for all involved: on the other hand, it is conducive to this type of government that the parties act as unitary actors (De Winter, 2002). Another feature is that in most of these cases the dominance of government over parliament is limited. Hence, parties in government and in opposition tend to seek cooperation rather than conflict (Tsebelis, 1990; Keman, 1997; Lijphart, 1999).
The final category with respect to the fabric of party government concerns those cases where this unifying behavior is seemingly absent—for example, Belgium, Italy, Slovakia and Switzerland. For historical and special reasons it appears that cabinet government is better off without the restrictions of collegial behavior in these countries. In Switzerland central government can be considered as an executive committee held together by means of its ‘magic formula.’7 Belgium established this practice to enable coalitions across the lan-guage-cum-territorial divide (Keman, 2002a), whereas Italy developed it to allow for minority government as well as to exclude the Communist party (during the Cold War). Finally, in Slovakia no party government could otherwise have been formed due to the deep-seated divisions within parliament.
In conclusion, the organization of party government is quite diverse. Four types have been distinguished here: collegial, prime ministerial, ministerial, and dual cabinet government. It is obvious that collegial cabinet government is the least hierarchical of the four parliamentary forms of government. The prime ministerial and dual forms are more or less of a hybrid nature, whereas the ministerial form is most typical for a government dominated by parliamentary parties (De Winter, 2002). This section on the fabric of party government in representative democracies shows, inter alia, that its organization is in part shaped by formal rules and related conventions. And yet students of party government more often than not overlook precisely the role of formal and informal institutions.
Political Institutions Matter for Governing—but in what way?
Governing is not just about settling conflicting demands and opposed ideologies between parties. Politics is embedded in institutional arrangements that have been devised to process citizens’ preferences and collective demands and to manage these conflicting demands. This requires a certain consensus among political parties for further cooperation by avoiding enduring stalemate. Institutions are seen to modify and regulate the behavior of the political actors (Tsebelis, 1990; Shepsle, 1995; Scharpf, 1998; Keman, 1999). Institutional arrangements differ considerably across nations and in their degree of formality. They may be simply established practices to tackle problems (e.g. deadlock and stalemate situations) or they may be derived from constitutions and related basic laws that direct the process of decision-making (Weaver and Rockman, 1993; Colomer, 2002). Surprisingly enough, this feature of representative democracy is often overlooked or taken for granted in the literature on the ‘partisan theory of policy-making.’ This points to a bias which should be avoided: actors alone are not sufficient to explain the organization and working of party government. Hence, we argue that the set of institutions and practices of a polity may well influence the room to maneuver of party government. It would be wrong therefore not to include the institutional configuration of representative democracies if one studies the question of whether or not party government matters with respect to the crucial purpose of governing: solving the problem of collective action by means of public authority. And this is dependent on the institutional arrangement within which a party government operates. In the literature one finds many different strands of thought that discuss this approach. We confine ourselves here to the latest book of Arend Lijphart, Patterns of Democracy, in which he has made a seminal contribution to the study of the relation between the institutions and public action of party government.
Lijphart (1999) focuses on the institutional constraints that can be derived from basic laws. In addition, the focus is on the institutionalized practices that have developed as consequences of formal requirements (see also Weaver and Rockman, 1993; Czada et al., 1998; Lane and Ersson, 2000; Colomer, 2002). The study of institutions is then considered as crucial to understanding the public actions of party government. The point of departure is the proposition that the institutional configuration of the decision-making process in relation to political actors and organized interests in representative democracies drives this process.
Lijphart (1984) developed a dichotomy with respect to the working of liberal democracies: consensus democracy versus majoritarian democracy. In Patterns of Democracy he developed both types further, and related these types to the way each performs in terms of governing the realm. Lijphart claims that consensual types of democracy perform better (even if they are necessarily not ‘kinder and gentler’) in the sense that they are more responsive in translating citizens’ preferences into governmental action by providing higher levels of public welfare. Why is this?
The first point is that when a democratic polity is forced to make decisions under circumstances of divided social constituencies then the institutional context must be organized in such a way that veto players tend to comply and cooperate. This would solve the so-called ‘collective action’ problem (Keman, 1997; Scharpf, 1998). Therefore parties (and organized interests) ought to be allowed to have access to the decision-making arena, on the one hand, and must all be in position to gain from the eventual results, on the other. Of course, this requires that the alternative option of maximizing a single party’s own gains is precluded. This is precisely what often occurs (sooner or later) within party democracy where conflicting demands prevail, and where the institutional context allows for stalemates (see also Tsebelis, 1990; Lane and Ersson, 2000). According to Lijphart, the principal requirements of a consensus democracy are:
- Broad coalition government which induces power sharing and may well be conducive to optimal policy choices by parties ingovernment;
- Relative dominance of government over parliament which allows for discrete policy choices on the basis of ‘government agreements’ enabling policy viablecoalitions;
- Proportional representation, allowing for all ‘minorities’ to be represented in parliament which may more or less diminish the use of veto play.
In short, consensual democracy is an institutional configuration allowing for political compromise and viable governing by party government. At the end of the day—according to Lijphart—this type of representative democracy will perform better. This is less likely to occur in a majoritarian type of democracy where confrontation and the principle of ‘winner takes all’ prevail.
The second line of thought regarding the impact of institutions concerns the way societal interests are represented with respect to the scope of political decision-making. In political science the idea of interest representation—parallel to that of parties—and, in particular interest intermediation is often called corporatism (see Schmitter and Lehmbruch, 1979; Crépaz, 1992; Woldendorp, 1997). Corporatism is the emergence and eventual institutionalization of consensual relations between organizations representing labor, capital (or business), and government. According to students of corporatism this institutional mechanism has been a key feature of concerted policy-making, particularly in many Western European democracies, in the post-war era (Armingeon, 2002). The basic idea is that it is possible to transform conflicting societal interests from a zero-sum to a positive-sum game. Most students of corporatism have pointed to the beneficial influence of this institutionalized pattern of interest intermediation. However, a requirement is that party government be composed in such a way that the corporatist forms of interest intermediation are recognized (Keman, 1999; Armingeon, 1999). In other words, corporatism can be seen as an institutional configuration devised to promote policy concertation and to foster political co-operation between societal interests and party. It will be clear that corporatist institutions will only bear fruit if and when party government is able to develop policies that foster stable policy outcomes.
Both institutional arrangements, ‘consensus democracy’ and ‘corporatist intermediation,’ are considered as important assets to create consensus rather than conflict in representative democracies. Institutions are considered as constraints on party behavior in government, but also as an opportunity structure for parties to further their own interests (Shepsle, 1995; Scharpf, 1998). Comparative research has demonstrated that these institutional arrangements shape the room for maneuver of party government (Katzenstein, 1985; Scharpf, 1987; Laver and Shepsle, 1996; Woldendorp, 1997; Lane and Ersson, 2000). Interestingly, the results are interpreted differently. A number of studies conclude, like Lijphart, that institutional configuration favoring consensual behavior of parties and organized interests is better for party government and society as a whole. Conversely, other studies do not deny the impact of institutions and related behavior of party governments, but rather question whether or not this itself enhances the democratic quality of governance. In this view, political parties, organized interests and the rules of the game within a democratic polity not only shape the room to maneuver, but also define the quality of governing by party government in a representative democracy. That quality depends on the degree of ‘responsiveness’ of parties and on the ‘accountability’ of party government.
Party Government: Responsiveness and Accountability
Responsiveness and accountability are two important aspects of the process of democratic decision-making and therefore central features for assessing the quality of any party government. Responsiveness—the extent to which parties in and out of government do indeed translate citizens’ preferences into public policy choices—is reflected in the relationship between ideological position and policy stance of parties. Accountability is the extent to which parties in government do indeed carry through their policy promises made during election campaigns. This Schumpeterian view of democratic quality still remains a valuable idea in assessing whether or not representative government works well (Klingemann et al., 1994; Budge et al., 2002).
In this section we will focus on the responsiveness and accountability of parties and government, an argument that is strongly linked to the mandate theory. This theory assumes that voters expect parties to fulfill their promises once they are in office, i.e. participate in party government. According to the mandate theory, the way parties govern depends not only on their ideological or programmatic stance, but also on the type and form of party government. This implies that the ‘mandate’ will not and often cannot be carried out as originally was thought. Government agreements, unstable coalitions or minority government and changing (external) circumstances often stand in the way. Yet, a number of students of government have attempted to establish how responsive parties are, in particular when they are in government, and to what extent party governments are accountable for their governing.
An empirical approach to testing mandate theory has been developed by the Manifesto Research Group (Budge et al., 2001). On the basis of a comparative contents analysis of party programs in OECD countries, it was possible to describe which political parties emphasized what salient issues in terms of policy priorities. Hence, the degree of responsiveness could be established in two ways: first, by comparing issues with programs; and second, by comparing how parties responded in terms of policy choices (see Laver and Budge, 1992; Pennings et al., 2005).
From the analysis it became clear that parties in general do respond to changing situations -for instance, through international crises and in the domestic economy (Keman, 1993; Pennings, 1997). Surprisingly perhaps, the response is not very different among parties. Yet it becomes different if and when parties are in government. Although parties do react to external stimuli, such as political issues and socioeconomic problems, party government actions appear to depend on its type and color (see Laver and Shepsle, 1996; Keman, 2002b): the more leftwing a coalition is, for instance, the more active policy-making will be if levels of unemployment rise. Conversely, the more rightwing a coalition is, the less generous the welfare state tends to become (Castles, 1998). It also makes a difference whether a government is ‘single-party’ or an ‘oversized coalition’: one-party governments tend to make drastic policy changes, whereas (broad) coalitions appear to go for ‘piecemeal engineering’ (Gallagher et al., 2005; Royed, 1996).
All in all, one may well conclude that parties are responsive to what goes on in society and can be held accountable for how they react if and when in government. At the same time, however, it is also clear that it depends on the type and color of government, and that its form also matters.
Conclusion: Party Government Matters!
Party government—making policy choices, allocating fiscal means, and producing rules—is one of the most discussed items in the media. The general public assumes that policy-making by government is by and large a response to their demands and related preferences. In a representative democracy, party government is held responsible for managing and solving collective action problems within a society. Not only symbolic responsiveness counts, but also the material output that is produced. Government, and in particular the parties participating in it, are held accountable. This implies that the academic debate revolving around the question of whether or not ‘parties matter’ has wider implications. Answers to this question also imply an assessment of the working of representative democracy and whether or not political parties are functioning in an adequate fashion (Dahl, 2000; Schmidt, 2000; Keman, 2002c).
Our analysis has revealed that, generally speaking, party government influences the type and direction of policy choices made in a representative democracy. This conclusion holds notwithstanding the fact that other factors—economic circumstances, demographic development and path-dependent trajectories—are relevant as well. Hence, it can be argued that the ‘partisan theory of policy-making’ is a viable approach to understanding party government. However, one must bear in mind that the formation of a government and the resulting form are complex features of party government that influence its room to maneuver. Equally obvious is that the institutional make-up of the government and concomitant mechanisms that allow for more or less consensus, co-operation, and viable coalitions are also conducive to good governance. Precisely for this reason, it makes sense to investigate the quality of this process in terms of responsiveness and accountability in a representative democracy.
The overall conclusion is therefore that the institutional organization of representative democracy is a determinant of the capabilities of parties in government with respect to collective action problems that vary according to the extent to which consensus and integrative mechanisms exist and do work. The ‘proof of the pudding’ of this lies—so we argue—in the extent to which party governments are ‘responsive’ and ‘accountable,’ on the one hand, and in the way political parties are able to function properly as viable and credible agents between politics and society, on the other hand.